The conjugacy problem in cyclic extensions of one-ended hyperbolic groups

Abstract

Let $G$ be a torsion-free one-ended hyperbolic group and let $\varphi \in \mathrm{Aut}(G)$. We prove that the mapping torus, or suspension, $$M:=G_{\varphi }=G{⋊}_{\varphi }\mathbb{Z}$$ has solvable conjugacy problem. This builds on the pioneering work of Préaux who solved the conjugacy problem for all (geometrisable) three-manifolds. We view $M$ as a generalisation of a fibred three-manifold. Our proof begins with the canonical JSJ tree of $G$, whose suspension gives a graph-of-groups decomposition of $M$. We refine each suspended QH vertex using a Nielsen--Thurston reduction system for its induced monodromy, taking account of non-orientable surfaces and orientation-reversing monodromy. Equivalently, this is a further geometric JSJ decomposition for each vertex which is a fibred three-manifold, but allowing Klein bottles as well as tori in the splitting. We then `block' pieces of this refined JSJ together according to whether they share a central element, glued along an elementary vertex, by a sequence of folding operations. This new splitting is cocompact and acylindrical; in particular certain local `solution sets' turn out to be rational subsets of virtually abelian subgroups - the virtually abelian groups in question are products of the edge groups. Our version of Préaux's graph-of-groups argument then reduces global conjugacy to effective intersection and non-emptiness for these rational sets.

Updated to show that the block tree is automorphism invariant